—The generalizability of currently available estimates of survival after radical prostatectomy is theoretically limited.
—To obtain generalizable estimates of survival after radical prostatectomy.
—A population-based retrospective cohort study.
—Nine regions of the United States.
—Patients who were diagnosed with prostate cancer between 1983 and 1987 and underwent radical prostatectomy and lymph node dissection.
Main Outcome Measures.
—Proportional hazards models incorporating geographical region, age, race, pathological stage, lymph node involvement, and tumor grade to identify independent correlates of disease-specific and overall survival and life table analyses to estimate 10-year survival distributions.
—A total of 3626 patients with a mean age of 65 years were included in the study; 92.6% were white, 54.2% had moderate-grade cancer, 60.4% had no extension beyond the prostate, and 91.2% had no lymph node involvement. Using San Francisco—Oakland, Calif, as a reference region, no other region was significantly associated with a risk of disease-specific or overall mortality. Older age and black race were independently associated with worse overall but not diseasespecific survival. Higher grade, extension beyond the prostate, and lymph node involvement were independently associated with worse disease-specific and overall survival. Estimates of 10-year disease-specific survival ranged from 75% to 97% for patients with well-differentiated and moderately differentiated cancers and from 60% to 86% for patients with poorly differentiated cancers.
—Neither disease-specific nor overall survival varied by region, suggesting geographically uniform assessments of risk in patient selection for radical prostatectomy. Across regions, overall survival varied by patient and prostate cancer characteristics while disease-specific survival varied substantially by prostate cancer but not patient characteristics. The present analyses provide the most generalizable current estimates of survival after radical prostatectomy.